I bought me an AD9833 Need a tad of advise

[Ian Rogers]

I am replacing my "home made" bench top Frequency generator!

This lil device can give a damn good sine at 10Mhz so I'm a bit chuffed.
However!! ( always a glitch) The sine and triangle pk to pk is only 0.65mV as per datasheet so I'm okay there.
The square however bypasses the inbuilt dac and is 5V pk to pk on the same output!!! so I need a decent opamp to give me a variable amplified output ( I found a decent rail to rail one that'll see 44Mhz.. but I'm going to need to switch the square wave in and out BUT!!!! then I won't have an adjustable square wave amplitude! so I need to attenuate the square before the amp and also switch it in.... Any idea's

 

[tepalia02]

Here is a design that you may take a look at. https://www.pcbway.com/project/shar...l_three_signal_generator_based_on_AD9833.html

This project is a single-channel digital signal generator designed based on ADI's AD9833 digital signal direct synthesis IC design, supporting sine, square wave, triangle wave

And you can use two potentiometers to add DC offset or adjust amplitude, respectively.

 

[Ian Rogers]

Thanks for your input but!

I have a decent design... I already have the three outputs, but the sine and triangle are minute compared to the square wave.... I want all three to be the same... I have to use an opamp to attenuate the square as discreet components distort the square wave... Messing with analogue signals isn't my forte so I ask here..

So far I have two opamps one to amplify all three signals, one to mess the square wave up, and a 4052 to majigger the signals to work..But each IC brings cak into the mix...

 

[JimB]

Ian

I have done a lot of work with AD9850/1 and other DDS chips, but never used the AD9833.

Your comment:

The sine and triangle pk to pk is only 0.65mV as per datasheet

struck me as a bit odd, and so I looked up the datasheet on the interwebs and found this:

The DAC is configured for single-ended operation. An external load resistor is not required because the device has a 200 Ω resistor on-board. The DAC generates an output voltage of typically 0.6 V p-p.

Which is a larger output than you are seeing.

Question:
Did you build your own development board for the AD9833, or did you use a handy module from China?

JimB

 

[Nigel Goodwin]

I presumed it was a typo?, and he meant 0.65V or 650mV?

It's the silly square wave level that messes everything up - here's a site which 'might' help give a few ideas?

DIY AD9833 signal generator - first steps (part 1) - Daumemo

DIY waveform generator based on a AD9833 chip. First step - to design and test analog gain control circuit with a VCA822 variable gain op-amp.

daumemo.com

And this next one looked promising, but while he left space to add an amplifier, it seems he never managed it?.

I too have an AD933, on a Chinese module - built up on veroboard, along with an Arduino Pro Mini (for it's small size), a TFT display, and two rotary encoders.

I spent ages trying to get the AD9333 to work with a PIC, and couldn't seem to manage it - so I tried an Arduino, that didn't work either. Tried a different AD9333 module - both then worked!!!! - might have been an idea to try a different module many days before I eventually did.

Just to confirm - I never got round to adding an amplifier either, or indeed actually using the AD9333.

 

[Ian Rogers]

Yup.. typo sorry... 0.65v pk to pk.... Anywho!!! I had a thought.. on the schematic that tepalia02 kindly linked... I didn't realize on first viewing that the square wave "will" attenuate when the gain pot goes below 2.2K

Nigel Goodwin The chip worked beautifully... As I said.. 50 hz -> 5Mhz and the sine is super stable..
When it gets above 500Khz, the output voltage tails off a bit, but I can use it..

Jimbo!! Yes a chineseium model..

 

[rjenkinsgb]

Could you just add a diode clipper to limit the amplitude and set the gain so the sine wave is just below the clipping level?

 

[Ian Rogers]

Could you just add a diode clipper to limit the amplitude and set the gain so the sine wave is just below the clipping level?

I had thought of that but the lil resistive attenuator screws every thing up... This is the same square wave that drives the internal dac.. so EVERYTHING messes up..... Thanks for your interest.. I think I'll just do what the guy did on the PCB way site...

 

[rjenkinsgb]

Which module is it? There seem to be quite a few variations with rather different connections or onboard extras?

 

[Ian Rogers]

Which module is it? There seem to be quite a few variations with rather different connections or onboard extras?

The cheapest... I didn't want the coaxial connector or the amp... It just has the DDS chip and 24Mhz crystal.

I am replacing my old 3038 function generator I made back in the 90's "Good ol' EPE mag" It has always been cak as the caps are switched in and the noise was awfull.. Also the sine wave was rubbish from the off!!

I'll do a vid once I get the final stage on..... BTW Nigel Goodwin Most of the Arduino sketches don't work with this... I found a decent library and have done my own..

 

[Dick Cappels]

I find the jitter on the output of the AD9833 operating at a few tens of kHz to be a problem for me. In the application into which I designed it I added an LC filter (driven by the square wave output) to clean it up. Because of the jitter I did not follow through on the idea of making it into a signal source for the benchtop. According to literature on the Analog Devices site the jitter is inherent in the design.

One of the DDS chips with a much higher clock frequency would provide a much cleaner waveform.

By the way, the board I built was hand-wired using point-to-point wiring without a ground plane. It was difficult to get the '9833 working properly until I put a 220 0hm resistor in series with the clock line at the clock input to the module to damp ringing.

 

[Ian Rogers]

Once I have it done I'll check the signal from x to xxxxx and take a looksee... The main use is to test a Juntion Box we make and I put a 10Khz signal through, so I'm not overly bothered about a tad of jitter, as long as the jitter is the same in as out! ...

 

[Ian Rogers]

Okay.. Tiny update.

I have still used the AD9833 But now with an Arduino mini pro, with rotary switches + an ILI9341 TFT display.

Seems to work extremely well... I have an LM318 on the output and I attenuate the 5v square by grounding a potential divider using a spare pin.. All three signals enter the amp at the same level..

Very pleased with it... I'll get it in the box and see what it looks like.

 

[JimB]

Sounds good.

Always nice to hear a success story.

JimB

 

[Ian Rogers]

Yeah! I have shown it to Stu, who would have finished it for me..
I have a slight colour deficiency ( as most men ) and it appeared that the course freq adjust wasn't working
But the code I pilfered used light green numbering and a yellow index, well I couldn't see what was happening, so my son was called to watch the screen... Now changed it to light green and a nice rosy red for indexing..

Ain't life a ball ache...

 

[Musicmanager]

Hi Guys
In preparation for some experiments on this tack when my parts archive has the right bits in stock I've done quite some reading today and there is a considered opinion that many of the problems within AD9833 are solved with AD9850 or AD5851.
I've ordered all 3 for some play time so at the moment it is based on others opinions which are web based so may or may not be valid. I intend to build a basic 'rig' and then apply each purported theory to see what I can or cannot achieve .. .. .. .
Yes Jim, I'm am a fully signed up local lunatic !

MM

 

[Ian Rogers]

Plus 1... I think your a loony as well

 

[JimB]

A couple of years ago I bought (at reasonable expense) a Chinese evaluation board for an AD9910 DDS.

I could not get it to work, the most helpful words that I found on the interwebs said that Chinese AD9910 evaluation modules were junk and did not work well.
One hero went on to describe in great detail how he designed a new board for the thing.

Gloom.
So I put the thing in the parts archive and moved on.

A few weeks ago, I had the idea to revisit the AD9910.
Another interwebs search found something on the AD website, a basic "write to these registers and it will work" type guide.
So I did, and guess what, it worked in a reasonable manner.
Modified rapture!

The AD9910 has lots of registers to make it do all sorts of weird and wonderful stuff, and my original attempt at programming the thing was way off and never stood a chance of working.

Oh what fun I am having.

JimB

 

[Dick Cappels]

A quick tip: I added an AD9833 (one of those inexpensive modules one can get via eBay or similar) to an existing hand-wired pad-per-hole circuit board and had a terrible time getting it to work as the datasheet said. Then I put a 10k resistor in series with the data clock input, which killed the ringing on the line and made things work perfectly. This was acceptable even though it limited the maximum clock rate because the clock signals and other inputs were being bit-banged by a 20 MHz AVR controller. A careful PCB layout is the best solution of sometimes you can't stop and re-do everything.

 

[danadak]

A quick tip: I added an AD9833 (one of those inexpensive modules one can get via eBay or similar) to an existing hand-wired pad-per-hole circuit board and had a terrible time getting it to work as the datasheet said. Then I put a 10k resistor in series with the data clock input, which killed the ringing on the line and made things work perfectly. This was acceptable even though it limited the maximum clock rate because the clock signals and other inputs were being bit-banged by a 20 MHz AVR controller. A careful PCB layout is the best solution of sometimes you can't stop and re-do everything.

A 10K resistor or 10 ohms ?

Regards, Dana.